Harnad's response to Buck (4-30-99)

Dear Anne and coworkers in the CalTech Proposal for the Scholars Forum:

First, let me make it explicit, as I have done before, that the CalTech initiative is very welcome and should be supported. Let us in no way balkanize our efforts over minor local differences. The overall idea is excellent, timely, and should be pursued with energy and speed. The recommendations and criticism below are made in the interest of strengthening the Proposal by streamlining it, paring off what is not relevant or tenable.

You have found the right word for it: "model." Models are welcome when we are testing hypotheses. There are many hypotheses about how we might reform the peer review system. Let them be modeled and tested. But let us not link the fate of these models, their tests, and their outcomes, to something that is already true, and demonstrated:

The current refereed journal literature, as it stands, is an extremely valuable resource. It is what all future work is built upon. Perhaps a better form of peer review could make it even better, but that is an empirical question. What is undeniable is that the current refereed journal literature, such as it is, is and has always been an invaluable resource to researchers.

Currently, access to that literature is blocked by tolls for Subscription/Site-License/Pay-Per-View (S/L/P), even though the authors of that literature have given it away for free.

The cost of access to that literature, and the even greater cost of lack of access to that literature, are the reasons behind initiatives like the Scholars Forum. The objective is to make this resource accessible for free for all, online.

This objective has already been achieved in one particular sector of the literature, in Physics, by the Los Alamos Eprint Archive (LANL).


The acceptance and indeed dependence of the world physics community on

LANL today is the proof -- if proof was needed -- that a free Archive of the journal literature is indeed what the learned community needs and wants.

It follows that LANL is not a "model," but a proven, working principle. The gist of my recommendation to the Scholar's Forum is that that is the principle it should implement, not any untested model for the future of journals or their successors. The LANL principle has face validity, overwhelming proof in practice, and is ready for extension to all disciplines. Why take something we KNOW works stupendously well and weigh it down with unproven "add-ons" that could well prevent it from getting off the ground just when it's ready to carry the other disciplines skyward, just as it has done in Physics?

No one needs to transfer anything from anywhere. If there is a robust, reliable Pandisciplinary Archive, its ubiquity and perpetuity soundly underpinned by Consortia, then authors can self-archive both their own unrefereed preprints and their own refereed reprints, as they do in LANL, and, in future, journals (established and new, with classical peer review and experimental) can have "overlays" on the Forum, as they will on LANL, in which they can officially certify the refereed papers. But for now, author self-archiving of preprints and reprints is enough, as it was in LANL. The rest will take care of itself, once the LANL revolution has spread to all the disciplines.

There is a crucial misunderstanding here: Self-archiving means archiving by the author. This is not restricted to local self-archiving in the author's Institution's Home Server (although that is an important component in creating a safely distributed, redundant and mirrored storage architecture). Authors need to self-archive in local as well as global Archives, such as LANL (and, one hopes, Scholar's Forum), with all the consortial resources you mention, to ensure the reliability and preservation of the global archive.

Once the literature is indeed all being self-archived, the Online medium itself will offer a wealth of resources for safely storing it, upgrading it with new technology, preserving it, and so on. The hurdle to cross is creating a LANL-style Archive in which authors are indeed motivated to self-archive all their papers in the first place. That's the hard part. The rest is implementational details. But if you needlessly overstructure the simple target -- a LANL-style Archive in which all authors place their preprints and reprints -- weighing it down with further, unproved structures, instead of restricting it to the tested and proven structures of LANL, you are needlessly jeopardizing the prospects of getting the literature online and free, as it has become in LANL.

If something has worked so well, why not extend it as is, instead of trying to "fix" it first, especially when all the "fixes" are controversial (and some of them incoherent), and hence likely to inspire needless opposition?

I don't know what "duplicative journals" means, and no one wants to put publishers out of business. The objective is only to free the refereed journal literature, online.

Societies, conferences, journals, etc. already have standards and protocols, and these are not uniform, nor should they be; nor is there any reason for them to want to submit their standards and protocols to any other body: This is not where the problem lies! Let standards and protocols take care of themselves -- or let CalTech establish a second, parallel initiative for the reform of standards and protocols. But for the present initiative, focus on freeing the literature, such as it is, standards and all!

Nor is founding new online journals a particularly useful goal. Let new journals form as they are needed, as they have always done. The goal here is to get the existing journal literature online and for free.

> a.) The Consortium will not support any S/L/P terms from anyone for access to
> the materials on the Consortium servers.

A worthy and desirable outcome, but COMPLETELY unrealistic under these conditions: Suppose I publish a successful paper journal now. I am interested in producing an online version of it too. What should I do? Set one up myself (as most journals are indeed doing) and continue to sell it, like the paper version, for S/L/P, or give it to the Scholar's Forum for free, so they can in turn give it away for free?

There is absolutely no incentive at all for a publisher to seek or accept such a "service"! But with just one slight parametric variation, almost exactly the same thing DOES make sense: if it is AUTHORS who are offered the free server, in which they can make their papers accessible for free for all. Journal authors already give their intellectual property to their publishers for free, to sell via S/L/P; they can in addition put it in the Scholar's Forum, where it can be accessed online without any barriers from S/L/P.

Vide supra. Why on earth should a publisher want to accept this? [Consortium terms for server access.] It is authors who can and will.

> c.) Authors or their institutions retain copyright per the terms established by the
> Consortium (see the Model text).

Agreed. Extremely important for their right to self-archive!

> 4. The journal as a finite aggregation of vetted works will fade away.

Yes, this will indeed happen, but only if the self-archiving (subversive) route is taken, as in LANL. If it awaits prior acquiescence by publishers, it has a long wait! (If it hopes to entice authors away from its proven, prestigious, high-impact journals to new journals or whatever, it has an even longer wait!)


> Thus, a society may announce the availability of a work of particular interest to
> members but that may be outside their usual focus or discipline.

As I said before, this is so trivial and out-of-date for the online medium that it is not worth spending any time thinking about.

> In addition, this platform will enable individuals to establish their own subject
> profiles to create personal alerting services.

Also irrelevant and trivial. Why speak of such short-sighted banalities in the same breath as the revolutionary possibility of freeing the entire journal literature online?

> 5. The Forum provides flexibility to authors:

Again, a banality. Authors always have the choice about whether or not to self-archive their preprint publicly before refereeing. I am afraid that if you keep dwelling on this sort of empty nonsense you risk not having the proposal taken seriously at all.

> Authors do not have to relinquish their right to submit papers to established
> journals; the Forum represents one option for authors.

This is completely incoherent. It is not even clear what this "option" is, once one clears away the fog: The Forum is just an Archive. It will contain either unrefereed preprints or refereed reprints. But to get from unrefereed preprint to refereed reprint one still has to submit the paper to a journal (established or new, conventional or experimental) for refereeing. So what's new here?

If the Forum does not support any new forms of peer review (protocols? standards?) of its own: that's very good. Then it is even clearer that it is just an Archive (although "just" is an understatement, for a free, global archive like LANL is just what Learned Inquiry needs).

But then why mention peer review at all? The Forum does not want to take it over, and become a megajournal or journals: good. It dreams of an arrangement whereby publishers seek to archive their contents in the Forum, offering them free for all. Dream on. But AUTHORS will certainly want to do this, once the Forum is available to them. So do so, and say so!

> Editorial Boards will continue to be established just as today.

This passage has absolutely no content. It is like saying that life will go on: Of course peer review will proceed apace. Why even mention it, or link it in any way with the substantive thing the Forum is trying to do?

> 7. Copy-editing is a minor process.

I agree, but its value is still nonzero, and will be part of quality control, along with peer review, no matter who implements it.

> …demand that any author have a work copy-edited to meet its quality standards.

Such decrees are, I regret, ludicrous! Why is the Forum pronouncing on all these things that are not, will not be, and should not be under its control or aegis at all? Implement a free archive and let the user community take care of the rest, as it did with LANL!

> The Forum proposes to eliminate the need for an editorial board to contract with
> or become a publisher to accomplish copy-editing…

It is beyond me why any of this is even being mentioned, but if not dropped, it will risk giving a potentially serious and substantive proposal an air of amateurism and even quackery.

I've commented on this once already: Distinguish pre-refereeing help in writing from post-refereeing help. The latter is called copy editing. No one can do it entirely for himself. Your friends can't be counted on to do it for you, for everything, always. Even a departmental service may not meet the standards of Journal X. So it follows that part of peer review, if you like, is some quality control for FORM, along with the usual quality control for CONTENT.

The rest is again amateurish future-casting (and does not sound very informed by prior editorial office experience either!)

Please forgive the irony, but it is so disheartening to see really good stuff mixed in with nonsense like this! Just drop it and the proposal will fly!

Yes, yes, conference proceedings (unrefereed, or lightly refereed) were always a problematic case, and often no publisher wanted to do them (and no doubt many conferences never got published, for that reason). They are a natural for Archives, but they are the easy (and small) part. The refereed journal literature is the hard (and much larger) part.

LANL, by the way, has already implemented this for at least one conference I know of, with no fanfare, and without any implication that any new principles are involved: The unrefereed conference literature is virtually preprint literature anyway. Conference overlays, to add weight to such preprints in the form of another kind of certification tag on which users can search (unrefereed conference proceedings, refereed conference proceedings), will be a useful feature.

If you continue to be vague about Journals, the hard case, and verbose about conferences, the easy case, you risk turning the Forum into a conference server, and delaying still further the freeing of the all-important journal literature.

Commentaries, as I said before, are yet another matter -- interesting, important, but nothing the Forum needs to make any explicit commitments about. Comments -- both refereed and unrefereed -- are among the many items that authors might self-archive in the Forum (along with revised updates, corrigenda, and other potentially linked material).

> 10. The Forum proposes to develop a dynamic alternative to deadening email
> discourse surrounding works in servers.

The "Forum" will do nothing of the sort (and does not gain credibility from so self-proclaiming, any more than Al Gore does when he claims to have created the Internet!). This has already been done, via software, in the form of the many Variants of Hypermail Archives that are in growing use today. Such threaded, linked comments and responses are again candidate items for deposit in the Archive.

> …develop the mechanisms to support and retain linked commentary.

Eminently worthwhile feature, but already proposed by the American Physical Society in 1992, and as obvious a capability for an online archive as citation-linking.


> The author of a comment on a refereed work may also submit it directly to an
> appropriate editorial board for review and possible inclusion in the archival

Would this not be amply covered by a generic statement that the Archive, besides containing unrefereed preprints and refereed reprints of papers, will also contain refereed and unrefereed commentaries on those papers, with links?

Online search and retrieval tools are a growing research specialty. Unless the Forum somehow wants to become, or to claim to be, especially expert or advanced in this specialty, they are best leaving it to the experts, and focusing on the invaluable resource the Forum really IS in a position to provide: A free online archive. Don't hitch your wagon to any current vision of the shape that online search and retrieval will be taking in the near future.

Please, forget about models and paradigms! Apart from a few platitudes along with a few bloopers, which should all be dropped, we have here a proposal to extend the proven LANL principle to the rest of the disciplines. Go ahead and do it! It will work, and a free, globally accessible learned journal literature WILL produce a new paradigm for the use of the learned literature, to the eternal benefit of research and researchers. The rest of the visionary stuff only obscures the picture!

Best wishes

Stevan Harnad harnad@cogsci.soton.ac.uk

Professor of Cognitive Science harnad@princeton.edu

Department of Electronics and phone: +44 1703 592-582

Computer Science fax: +44 1703 592-865

University of Southampton http://www.cogsci.soton.ac.uk/~harnad/intpub.html

Highfield, Southampton http://www.princeton.edu/~harnad/intpub.html